Q-Day Conference 2023 showed that Generative AI is ‘doubly’ decodable
The Q-Day Conference took place on September 19 and brought together more than 1600 participants – those present at the Culturgest auditorium, in Lisbon, and those who followed the event online, via live streaming.
The 14th edition of the Q-Day Conference, the annual event organized by Quidgest, had the motto “Decoding Generative AI”. João Paulo Carvalho, Co-founder and Senior Partner at Quidgest, took advantage of the event’s Opening Session to explain the reason for this theme, saying that, on the one hand, Generative AI is decodable, because “it’s not magic, it’s not supernatural, it’s not inexplicable. It’s science, it’s a skill, it’s something we can all understand and use”, and on the other hand, “Generative AI automates software development, taking the focus off the code (decoding) and leaving more time for solutions”.
Recalling that Quidgest and the Genio platform were born with the second wave of AI and have always maintained their commitment and investment in R&D, during the decades of “Winter” that saw other players give up, João Paulo Carvalho stressed that the ‘generative’ component is at the origin of Genio itself… for 33 years now! The Quidgest co-founder also stressed the importance of holding the Q-Day Conference for 14 consecutive years, because “when we master the concepts, we dispel the dystopias about AI” and to delve deeper into this topic is to understand that “Artificial Intelligence is many things, but it is always Human Intelligence”.
The event, which was recognized with the high patronage of the President of the Portuguese Republic, Marcelo Rebelo de Sousa, showcased new horizons in technology and digital transformation, spread over four large thematic panels:
Panel 1 – Decoding Generative AI to Empower Positive Change with Global Purpose
João Paulo Carvalho returned to the stage as the speaker on the first panel, this time to focus on the differences in perspectives on Artificial Intelligence. If the distinction between Discriminative AI and Generative AI is peaceful (the former suggested content we would like to see, but couldn’t explain why; while the latter is capable of creating content and/or images we like, for example), the same is not true of another fundamental difference in perspectives in AI.
This is the difference between Symbolic AI (tries to replicate rational and conscious thought; models human reasoning through explicit rules and algorithms; aims to create systems that think logically, solve problems, demonstrate theorems, and are understood by humans) and Neural AI (tries to imitate the functioning of neurons in the human brain; is not concerned with explicit logic or conscious reasoning; focuses on creating systems that, without resorting to explanations, are effective at establishing statistical associations between large volumes of information).
To simplify and contextualize this distinction, João Paulo Carvalho said that Symbolic and Neuronal are the “how” of Artificial Intelligence. Discriminative and Generative are the “what for” of Artificial Intelligence. Neuronal Generative AI (ChatGPT) is very good at recognizing patterns in large sets of information. Symbolic Generative AI (Genio platform) is reliable, traceable, strong in logical reasoning, and explicit representation of knowledge. “Combining Neuronal and Symbolic Generative AI is the most promising bet”, especially when we apply this combination to software development at Quidgest, concluded the speaker.
Paulo Dimas, VP for Product Innovation at Unbabel and CEO of the Center for Responsible AI, began by challenging those present on the impact of Generative AI, which promises to be equivalent to 10X Portugal’s GDP in just 10 years, according to recent studies by McKinsey & Company. User growth is increasingly rapid, identifying the author (human or artificial) of texts and images is becoming almost impossible and, in the end, we’re talking about 300 million jobs being impacted.
But will AI destroy humanity? Maybe AI will save us from extinction. But there are serious risks, the speaker explained, stressing the importance of ensuring Responsible AI – AI that is both fair, explainable and sustainable. Paulo Dimas said that this is the mission of the Center for Responsible AI and that there are currently 21 products that use Responsible AI, in areas such as emergency room diagnostics, access to health data, or clinical research. “We are contributing to the UN’s Sustainable Development Goals”, he concluded.
Pedro Oliveira, Dean of Nova School of Business and Economics and Professor at Copenhagen Business School, explained in his presentation how AI can be used for Social Good, helping patients to help themselves.
The speaker used his presentation to explain how AI contributes to data interpretation (to better understand medical conditions and/or gaps); design assistance (designing and prototyping medical devices); personalization (so that devices or software are better adapted to patients’ individual needs); rapid prototyping (for rapid testing and speeding up the overall development process); and cost-effectiveness (fewer costs associated with developing new medical solutions).
Through specific examples in which a human asks an AI machine (ChatGPT) for precise indications to solve concrete problems, such as asking for the “step by step to avoid being bitten by mosquitoes and infected with malaria”, it is possible to create products/repellents within anyone’s reach and thus give new meaning to the word innovation. “AI is going to totally revolutionize the way we think about innovation and entrepreneurship processes”, summed up Pedro Oliveira.
Sandro Mendonça, Professor at the Department of Economics, ISCTE Business School – Lisbon University Institute, brought the topic of “AI and Technical Standards” to the table, i.e. the layer of intangible infrastructures: role, progress and potential. Pointing out that Generative AI is a topic that is plaguing the world and that technology itself is “on the loose”, the speaker also highlighted the current race for rules in AI. The awakening to regulation not only has various approaches (supra-national impulses in Europe, transnational intentions such as the Hiroshima process, and national initiatives in China or Brazil); but different paradigms proliferate (oriented towards “protection”, “risk” and “domain”).
The speaker said that “we are witnessing a real dispute over regulatory leadership and industrial leadership”. Standards are shared rules for designing or measuring products and processes. Standardization for change, on the other hand, turns out to be a specific type of innovation: “technical standards” are institutional/non-technological innovation; and these can be reactive, responsive, and anticipatory, bringing with them an equal account of virtues (lower transaction costs and greater economies of scale and network effects) and vicissitudes (pro- and anti-competitive balance, trade-off, and technological locks, lock-in or path dependence).
In conclusion, Sandro Mendonça explained how the structure of Standardization works in Portugal, mentioning the National and Sectoral Standardization Bodies, Technical Commissions, and Voting Members that currently contribute to the framework for Standardization activity in Artificial Intelligence.
Panel 2 – Decoding Generative AI to Provide a New Generation of Public Services
César Pestana, President of the Board of Directors of the Portuguese Public Administration Shared Services Entity (ESPAP), opened the presentations of the second panel by talking about the potential of AI for streamlining, simplifying, and making processes and procedures more transparent in the Public Administration.
For the speaker, there are different factors that facilitate this reality: a context of significant digital (and climate) transformation with the support of the Recovery and Resilience Plan (RRP); a very important market of technology suppliers at national, European and global level; a clearer legislative framework in technological areas; and a very high culture of technological adoption, both in the public and private sectors.
In the specific case of ESPAP, the President of the Board said that the following initiatives are being carried out: collaboration protocols with universities to promote understanding of the theory and practice of AI and Machine Learning; AI in chatbots, applications, and portals; AI in RPA in technological support operations; but also the planning of its application in the renewal of transactional systems and data analytics, in most areas integrating information.
João Dias, President of AMA – the Administrative Modernization Agency, challenged the audience to reflect on the theme of “AI in Public Services for the Citizen”. The speaker reinforced the message that Portugal is at the forefront of digital transformation, namely through its presidency of the Digital Nations alliance, its improvements in the DESI (EU digitalization index), as well as its progress in the Open Data Maturity Report.
As examples of innovations in the Portuguese Public Administration, João Dias mentioned the complete dematerialization of the Citizen’s Card, the use of the cell phone as an identification document, the European recognition of the dematerialized Citizen’s Card, and the SNS24 App for accessing health test results.
Closing his talk, the speaker reminded the participants of AMA’s Mission: “Simplifying the lives of citizens”. How? By investing in digital transformation to improve public service – as exemplified by AMA’s virtual assistant, based on the GPT 3.5 Turbo model, and capable of recognizing voice and writing in Portuguese, synchronizing mouth movements, and generating audio, providing 24/7 service. “Artificial Intelligence is a fundamental ally in the search for more efficient and accessible public services, benefiting society as a whole”, concluded João Dias.
Pedro Dominguinhos, President of the RRP National Monitoring Committee, reminded participants of the nature of the RRP – a program based on a European mechanism to respond to the Pandemic, with a view to making society more resilient and promoting the dual transition: climate and digital.
In all, there are three dimensions covered (Resilience, Climate Transition, and Digital Transition), 20 components and 115 measures (32 reforms + 83 investments) totaling 16,644 million euros. But what are the opportunities of the RRP combined with AI? Pedro Dominguinhos explained that the RRP provides a set of opportunities/investments that aim to respond to various challenges: more competitiveness and innovation; more skills and better alignment with the challenges; and better, more digital and accessible public services.
In the case of companies, mobilizing agendas, companies 4.0, and business capitalization are expected. For citizens, the opportunities translate into qualifications and skills, companies 4.0 with digital training and training for the Public Administration. In the case of the Public Administration, the investments go towards the digital transition in health, the quality and sustainability of public finances, economic justice and the business environment, as well as a more efficient Public Administration. Before concluding, the speaker also stressed the “importance of partnerships in the ability to leverage investments and boost results and impacts”, as well as the “importance of companies having digital transformation plans in order to leverage the support available”.
Wagner Araujo, PhD, Research Associate at the United Nations University, Electronic Governance Unit, used his presentation to reflect on the purpose of Generative AI for the Public Sector. Using inspiring metaphors from the story “Alice in Wonderland”, the speaker explained that the objectives of an eGOV strategy are several: to make government more efficient; to make government more transparent and accountable; to improve public services; and to improve the relationship between the public sector and society. “AI should support the constant search for (good) governance? That’s what matters”, said the researcher.
But what is (good) governance? If the governance process goes well, it is usually referred to as “good governance”, with the government having the following characteristics: it is participatory and consensus-oriented; it is accountable and transparent; it responds to what communities need; it is effective and efficient; it is equitable and inclusive; it follows the so-called rule of law.
Using various examples to illustrate the point, the speaker also explained that “Generative AI has to be asked the right questions in order to provide an appropriate response”.
Marcelo Ribeiro, Software Developer at Quidgest, closed the second panel with a presentation on “The Value of Time in the Modern World”. The speaker recalled that technological proliferation is often associated with pressure for productivity, the need for spaces for reflection and introspection, and the importance of re-evaluating the true definition of a “productive day”. However, with the arrival of Large Language Models (LLM), it has become much easier to access large volumes of text, use advanced neural networks to understand and process data, and generate responses using coherent and contextual interactions.
For Marcelo Ribeiro, the impact of AI cuts across several domains and makes it possible to analyze requirements to more accurate estimates using historical data, automate and optimize the creation and maintenance of documents, help implement developments based on Natural Language Processing (NLP); create Chatbots that allow for a more natural and efficient interaction; or algorithms for the analysis of irregularities and patterns, allowing the user to focus their attention.
In the future, it is expected that AI models will continue to evolve and be trained for increasingly specific tasks, that data regulation and privacy will be a priority, and that companies that want to continue to thrive (and even exist) will prepare to integrate and embrace technological advances in this area quickly, concluded the Software Developer at Quidgest.
Co-Innovation Awards
The presentation of the Co-Innovation Awards was one of the highlights of the Q-Day Conference, as it recognized Quidgest’s clients and partners who participate most actively in the development of digital solutions that stand out in Portugal and around the world.
- Customers with History Co-innovation Award | ERC – Entidade Reguladora para a Comunicação Social;
- Customers with History Co-innovation Award | DROPEP – Direção Regional da Organização, Planeamento e Emprego Público do Governo Regional dos Açores;
- Science and Academy Co-innovation Award | Universidade do Minho;
- Digital Privacy Co-innovation Award | Grupo Lusíadas Saúde;
- Health Co-innovation Award | Centro Hospitalar Universitário Cova da Beira;
- Local Community Co-innovation Award | Câmara Municipal de Cascais;
- Qualification for the Creation of Public Value Co-innovation Award | ADENE – Agência para a Energia;
- Digital Transformation Co-innovation Award | Camões – Instituto da Cooperação e da Língua;
- Public Management Co-innovation Award | Direção Geral do Tesouro do Ministério das Finanças de Cabo Verde;
- Network Co-innovation Award | EEA Grants Portugal;
- International Partner Co-innovation Award | Kroll Portugal;
- Financial Sector Co-innovation Award | MarTrust Europe;
- New Most Valuable Product Co-Innovation Award | AGD – Accountant General’s Department, Jamaica
- Genio Partner Co-Innovation Award | Spacebar Tecnology Solutions.
Panel 3: Decoding Generative AI to Transform Business and Scale Success
Alina Cristina Buteica, L&D and HR Expert, Founder & CEO at Illuminated Essence & Growth Hives, spoke on the topic of “Conscious Development of Generative AI in Business”.
Challenging the audience with the question of whether we are at the peak… or just the beginning of Generative AI innovation, the speaker mentioned examples of businesses that are at the forefront of this area: Unbabel (helping companies deliver multilingual customer experience at scale) and Wild.AI (the Generative AI app that elevates the capabilities of the female body).
The entrepreneur also mentioned RedBridge Lisbon – the Entrepreneurial & Investors club bridging Lisbon & San Francisco, which has dedicated its activity to organizing events on AI & Business: Champalimaud Center, Lisbon; Shack15, San Francisco; RedBridge Palacio, Lisbon; TechCrunch Article after event with Unbabel.
Speaking about the case for Human-Led, Earth-Conscious (Generative) AI, Alina Buteica said that everyone should get involved in shaping the future of AI – we shouldn’t just leave AI to the experts (as it impacts our future). She also said that we should promote a responsible use of AI, promote compassion for the environment, human life, and all forms of life, as well as a biomimetic and ethical AI, which is at the service of our planet and human life.
“Nature without AI would still be a beautiful place… But AI without nature, living in a totally artificially generated/manufactured world… Is a frontier not to be crossed, for our greater good”, concluded the speaker from the third panel.
Filipe Marques, Chapter Lead at Data Analytics and AI at Siemens, brought to Q-Day an example of a successful case implemented in the company where he works: “ai:express – Automated Generative AI and Chatbot Deployment”.
Bearing in mind that at Siemens, “we aim to make LLM accessible to everyone, providing a platform that can be customized to cater to the unique needs of each business”, the speaker explained that this experience was born from the challenge of having a framework for domain adaptation across Siemens units, making it accessible to everyone.
In this way, ai: express emerged, with a user-friendly interface suitable for non-coders and security, integration with Siemens Cloud services, compatibility with various LLM models, and support for various file formats and URL parsing. What’s more, it is being tested for different use cases: Ai attack assistant, RolandUnplugged, SGES Topic Modelling, Conference Summarization, Integrated Management System Documentation, and Siemens Data Cloud, exemplified by the speaker on the third panel.
Hugo Cartaxeiro, Founder & Managing Partner at Singularity Digital Enterprise and Urbiwise – Property Markets Analytics, began by presenting a brief history of AI development, focusing on the differences between AI Parameters and the Human Brain, as well as the importance of Large Language Models: “LLM are unique in their ability to generate high-quality, coherent text that is often indistinguishable from that of a human”.
The speaker then addressed the topic of Generative AI Business Impact, saying that “Generative AI is set to unleash a powerful wave of productivity growth that is likely to affect all industries and could add up to 4.4 trillion dollars a year to the global economy”.
Regarding the importance of being a first mover, the entrepreneur said that these ‘champions’ are characterized by their competitiveness (early-access to cutting-edge research, tools, and platforms to create innovative solutions and products); impact (by solving some of the most pressing challenges and opportunities); growth (by automating and optimizing processes and also refreshing business culture with new ways of working and thinking); and learning capacity (different perspectives and experiences by expanding knowledge and skills).
Liliana Ferreira, Director of the Fraunhofer Research Center for Assistive Information and Communication Solutions in Portugal, dedicated his presentation to “The Future of Artificial Intelligence” topic.
Having previously explained why Artificial Intelligence is not Machine Learning, the speaker quoted Picasso – “Computers are useless. All they do is provide answers”, to clarify that “Generative AI is a tremendously powerful assistant. Humans have the responsibility to query the AI, and therefore to control its activity and to verify the relevance of its output”.
Liliana Ferreira recalled that AI currently does a number of tasks well: generating reports, supporting brainstorming, detecting anomalies or specific activities, segmenting information, translating texts, generating content from different data sources, “reading” images, creating transcripts, etc. But AI is not always used with good reasons, stressed the speaker, listing some cases of fake news, bots, trolls, or micro-targeting.
To conclude her participation in the third panel, Liliana Ferreira recalled the importance of projects such as the Center for Responsible AI and Artificial Intelligence Certification, which make it possible for data provenance and governance, traceability, accountability, and transparency.
João Ferro, Chief Technology Officer at Quidgest, began by saying that, “at Quidgest, we believe model-driven development shapes future-ready software” and “in the realm of Generative AI, models take center stage”. The speaker of the third panel, seized the moment to highlight some of the benefits of the Genio platform. The Quidgest platform is capable of automatic code generation, thanks to its model-driven development approach, resulting in ready-to-use solutions. Did you know that through the use of AI, Genio’s automatic code generation goes from the traditional human programmer rate of 3 bytes per second to a staggering 1,000,000 bytes per second? The benefits for business owners are clear: faster time to market, improved collaboration, and reduced development costs.
According to the speaker, there’s nothing to fear when it comes to traditional software developers losing their jobs: “Let Genio handle the repetitive part of coding, while you focus on making the valuable parts”. By using the creative capabilities of Generative AI and combining it with the predictability of an Inference Engine, we could tackle the biggest faults of both models and enable the creation of application models by prompting, instead of manual definition.
Panel 4: Decoding Generative AI to Reshape the Future of Work and Skills
Zoltan Istvan, AI expert and author of “The Transhumanist Wager”, joined us remotely to share a message about Transhumanism – a captivating and thought-provoking concept. This movement, which has gained substantial traction in recent years, envisions a future where the boundaries between technology and biology blur to unlock the full potential of human existence.
What’s happening to our world? “Over the past 12 months, Generative AI appears to have changed the world, arriving sooner than we expected”, stated the speaker. The former journalist also asserts that numerous professions will require reconfiguration. However, AI will also facilitate the exploration of additional medical treatment options and may even enhance our human capabilities by incorporating robotic components into our bodies.
Given this emerging technological and economic landscape, “AI is a new way of making money for many companies”, “a new motive to restart our professional careers”, but also “a new way of making companies better”, concluded the first speaker of the fourth panel.
Bernardo Ferreira, E-Commerce Marketing – SEO Specialist at Worten Portugal, took the stage to talk about “How to stay ahead in Digital Marketing”.
The speaker’s presentation highlighted several key points. First, he emphasized that we are living in a new search era, describing it as Google’s biggest transformation ever. This transformation has led to a significant change in customer behavior, with 46% of customer journeys now starting on Google. Consequently, this has disrupted traditional customer journeys, causing organic traffic to decline while directing more traffic to product detail pages.
Second, the speaker discussed the importance of visual content. He stressed that there is no longer room for “boring images” in the digital landscape. Next, he explored the potential for creating unique product descriptions using various tools like ChatGPT, Jasper, and Copy.ai.
Furthermore, Bernardo Ferreira touched upon the role of chatbots in enhancing customer engagement. By deploying conversational chatbots powered by Language Models (LLM), businesses can finally bridge the gap between automated interactions and the human touch, providing real-time responses to customer queries. Also in the context of the fourth panel, the speaker delved into the concept of personalization on steroids. This involves moving beyond traditional personalization to dynamic and generative personalization, with applications like virtual try-ons serving as notable examples of this trend.
Isabel Ramos, Associate Professor at the Department of Information Systems of the Universidade do Minho, chose “Synergizing Human and Artificial Attention: A Blueprint for 21st Century Organizational Excellence” as the theme of her presentation.
The speaker of the fourth panel explained that attention is “an integral part of countless mental operations – including understanding, memory, learning, reading emotions in other people – critical to the agility in life and excellence of what we do. Attention is a complex and fundamental cognitive human capability insofar as it influences perception, cognition, and behavior”.
Hence a wealth of information creates a poverty of attention, explained the speaker. Why? As of 2020, the world generated 2.5 quintillion bytes of data daily (SG Analytics, 2020). This figure is expected to grow exponentially in the coming years. More… With the proliferation of smartphones, social media, and the internet, people are exposed to an overwhelming amount of information and content. The constant notifications, alerts, and the “always-on” nature of our digital devices can disrupt our focus
But can we say that AI systems have the ability to pay attention? Isabel Ramos believes that research has conceptualized AI as the ability to select and prioritize input data relevant to the performance of a specific task. However, AI can be distracted or overloaded too, due to the processing of extended data sequences, the ambiguity or patterns lacking clarity within the data, the input data differing substantially from the data used for the AI system training, or the meticulously constructed inputs intended to deceive AI models.
Susanna Coghlan, Europe Business Development Manager at Quidgest, shared with the participants a presentation about the topic “Decoding Innovation: why Ideas matter more than Coding Skills”.
Taking a personal project in Cambodia as an example, the speaker explained that it is possible to run a small family foundation that is very well connected in the IT world thanks to donated laptops, logo – basic programming software, and Lego robotics kits. “The idea was that these tools are intuitive – just give them the kits and let them free to decide how they want to build and direct the robots”, explained the speaker.
Quidgest has a similar shared vision with Genio platform and Quidgest_Academy: equalizing opportunities by creating the ability for everyone to make an impact through digital transformation. “Because people in Quidgest believe that creativity should not be limited by technological skills. We believe people should focus on innovation and ideas for improving our world, not coding”, underlined Susanna Coghlan, referring to recent data from the Future of Jobs 2023 (World Economic Forum), while closing the fourth panel.
