Austin return (SXSW 2024)-Between promises and concerns: Tech’s growing reliance on AI
by
(Tech Community Advocate at )South By Southwest (SXSW) is one of the most important conferences for discussing current and future technological trends, and the 2024 edition lived up to all expectations. At Adevinta, we use our 25 digital marketplaces across 10 countries to help everything and everyone find a new purpose, whether that’s consumer goods, mobility, real estate, holiday rentals or jobs.
As a Tech Community Advocate at Adevinta, it’s part of my role to stay aware of technological, corporate and societal evolutions, and SXSW holds a special place for sensing the winds of change. I was lucky enough to attend this conference with three main objectives :
- Observing the main tech trends of this conference
- Meeting some inspiring GenAI leaders to compare their vision to our current experiments for our users and collaborators
- Connecting those thought leaders to my tech teams
I had the opportunity to meet Frédéric Josue, founder of the consultancy firm 18m, and an active observer of the constant evolution within our tech ecosystem. With over two decades of leadership experience in open innovation for major corporations, he has managed global R&D applied initiatives and provided strategic advice on cutting-edge technologies. He collaborated with top institutions like Technion, USC, UCLA, the University of Tokyo, KAIST and SNU, focusing mainly on AI strategies. As president of the consulting firm 18M.io, he specialises in open innovation and helps clients navigate complex, dynamic market challenges.
We noticed three messages that were repeated during the conference:
- The future is still a mystery
At the heart of this festival dedicated to new technologies, neither the industry giants nor the traditional companies seem to have a clear vision of the future. They all seem to sketch the future cautiously without being too detailed. - USA! USA! USA!
The 2024 conference also marked the predominance of the USA across all technological dimensions, from mastering hardware production with chips (Nvidia, AMD), to the cloud market, and even their computing power. The ‘Made in the USA’ message was loud and clear. - Ethical dilemmas
2023 was the year of all fantasies related to the development of generative AI. In 2024, AI ethics is at the centre of discussions, with a particular focus on algorithm transparency, corporate responsibility and the privacy issues associated with the increasing use of AI in various domains.
Upon our return from Austin, we set aside some time to discuss seven major trends highlighted during the numerous sessions of this conference.
1. DEI and Greentech, facing economic reality: How to support essential causes in a degraded context
In stark contrast to the 2023 edition, we felt that DEI topics were less prominent in the panels, especially in the face of the evolving global politico-economic context. This is most clearly marked by strong protectionist tendencies fuelled by a gradual West vs. East opposition.
The business world is digesting this climate of increased competitiveness by cutting budgets and resources dedicated to DEI and Greentech initiatives or by parallelising actions, thereby sometimes forgetting the intersection of identities and communities (women in tech, LGBTQI+, accessibility etc.). Users’ expectations often align with those of our collaborators. The development of ERGs (Employee Resource Groups) at Adevinta helps our teams better address the needs of our users. For example, the LGBTQI+ ERG has been able to assist our product teams in integrating the third gender option of “unspecified” into the user account details on our leboncoin platform.
=> Our food for thought is how we can adjust ESG/DEI approaches to political challenges and social imperatives?
Source: DEI in Tech: Restoring Commitments During Challenging Times
2. Mirror Effect and AI Privacy: Who should we now believe, human or machine?
The anthropomorphism of AI, which involves attributing human characteristics to technologies, can increase users’ acceptance and interaction with these systems. This increased familiarity with technologies may reduce individuals’ vigilance regarding their privacy. In other words, AI becomes a familiar entity of trust, similar to a human.
When interacting with digital assistants, users share a wealth of personal and behavioral data, raising security and privacy concerns. In the future, we can imagine that each consumer will be represented by personalized bots that will interact with brands and digital platforms.
=> Key takeaway is: how to design interfaces that preserve digital identities while offering relevant and reliable content — goods or services? And how not to lower our level of vigilance when AI is banally surrounding us in our daily lives?
Source: Accenture, Technology vision 2024
3. Undecipherable AI: Challenges posed by non-deterministic AI capable of acting without transparent explanation
Deep learning-based AI technologies introduce unprecedented complexity into algorithmic decision-making, often without the ability to trace the logical path leading to a specific conclusion. This opacity raises important ethical and practical questions, as it limits users’ ability to understand, challenge or validate the decisions made by these systems.
=> With the knowledge that, for example, 12 to 15% of Alphabet’s codebase is generated by machine learning, what mechanisms of transparency and accountability are needed for trustworthy AI? This is something we should continue to explore.
Source : Amy Webb, Emerging Tech Trend Report 2024
4. Truth in Question: The challenge of multimodal LLMs, between the risk of reality distortion or productivity opportunity
Multimodal LLMs, by integrating text, images and potentially sound, offer unprecedented interaction richness and immersion capabilities. However, this same ability to convincingly simulate realities can be used to manipulate perceptions and distort truths, posing a major challenge in terms of misinformation. The ease with which these models can generate credible and personalised content lowers the barriers for malicious actors, increasing the risk of large-scale influence operations. For example, we remember the enlightening talk by Paige Bailey (Lead Product Manager at Google Deepmind) illustrating Gemini’s numerous experimentation fields, citing the example of Gemini 1.5 demo capable of identifying errors in a codebase based on phone video captures of bugs in an application.
=> Something to think about is what strategies are there to affirm message integrity against misinformation?
Source: AI & Productivity: A Workplace Love Story?
5. GenAI market in full bull mode: Positive employment boom thanks to AI, need to anticipate and develop adapted training and a productivity increase without evidence
Adapting to an AI-dominated economy requires a cultural shift emphasising curiosity and continuous learning. The benefits in terms of productivity are not guaranteed and can vary considerably, highlighting the importance of employees’ adaptability. This transformation will lead to fantastic job creation requiring a range of new and varied skills. Patience and firm commitment from the organization are vital to succeed in this evolving environment.
On the tech side, support for engineering teams is already underway, for example, with the deployment of Github Copilot in more and more tech companies.
=> How can we implement a structural and, above all, cultural transformation for employees?
Source : Learn, Adapt, and Succeed in the Age of AI
6. Realpolitik in the face of common goods: Challenges posed by the possible weakening of open source
The difficulty in monetising Open Source leads startups to prioritize the protection of their innovations, turning away from knowledge sharing to seek more lucrative partnerships. This trend risks limiting access to innovative technologies and hindering the open collaboration that has long fueled advances in digital technologies. At the same time, Elon Musk — on the margin — opens the floodgates with Grok.
Ultimately, AI regulation is underway, and the Europe Union intends to play its part in defining the rules of the game.
=> Our key takeaway is how can we preserve open and sustainable digital spaces in service to society?
→ Source: Why the Future of AI Should be Open
7. Cognitive superpower: In the face of ethical challenges of technological solutionism, evolve behaviours or correct externalities through technology
Europe is betting on adapting human behaviours to address the ethical challenges of technology, while the United States prioritises solutions that support their capitalist interests, correcting externalities without altering consumption patterns. AMD — the second largest microprocessor manufacturer — has acknowledged that there is a supply issue with available natural resources mining, not a chip demand problem.
=> Food for thought: how can we design products and services that are more resource-frugal?
Source : Keynote: Lisa Su (Chair & CEO of AMD) in Conversation with Ryan Patel
Conclusion
SXSW 2024 was the year of an acknowledged awareness of the limits of AI: AI is uncertain, encouraging restraint — and a few less fantasies. Now, we assume it is conditioned by the evolution of our available resources and means of production, a discussion that continues among the GAFAM leaders. Generative AI is disrupting usage for the collaborator, the platform and the consumer. This growing technological anthropomorphism lowers our level of vigilance, and it will be imperative to establish governance and usage rules to:
- Strengthen alliances between AI publishers and platforms: we used to think that the key was cloud infrastructure and hosting our platforms. Now, CTOs have no choice but to establish collaborative partnerships with GenAI publishers and microprocessor producers to stay close to technological developments.
- Clarify capabilities and technical limits: it would not be surprising in the future for platforms to frame the technological stacks used and make them accessible to users, in the interest of transparency/accessibility.
- Promote ethical and responsible use: Encourage responsible use of technologies by highlighting their complementary rather than substitutive role in human relationships, particularly in areas requiring empathy and emotional support. Emphasise the adoption of a clear ethical framework that respects users’ privacy and dignity, while highlighting the brand’s efforts to ensure these principles.
SXSW 2024 and GenAI have once again caused us to question our company project at Adevinta, and are already guiding our work with Product & Tech, HR and communication teams. E-commerce will be even more conversational, and we have a responsibility to better understand and support the expectations of our employees and customers with technological solutions, training and clear and transparent communication about the adoption of AI technologies.
Being a leading online classifieds group, operating 25+ digital marketplaces across Europe and beyond, with +120 million unique users/month, we are looking at three main aspects to step up our AI usage: how users will find our marketplaces (e.g. search engines, AI assistants etc), how users will use our marketplaces (e.g. improvements in the user experience) and how our workforce build our marketplaces (e.g. productivity enhancements).
Overall, Europe is actively shaping its AI landscape with a focus on responsible and human-centric AI development. Regulations, investments, corporate innovation and societal dialogue will play crucial roles in shaping the continent’s approach to AI. To be continued…
