2024 will be about blurring lines between AI and human. It might be too soon for the “AI” word to fade as a buzzword, but what it enables embeds itself in every facet of what we do.
World
- Aging & care shape the relations between people.
- Climate change still progresses - and impacts everything else. Should I say more?
- Shortage on materials and weak supply chain links help steer reusability, reparability, and service-based initiatives, as opposed to selling stuff.
- Uncertainty forces insurers to find new ways or working.
- Uncertainty and agents drive polarization forward, weakening global institutions. Democracy suffer and inequality rise.
- The knowledge society - albeit a fraction of global society - redirects its attention to non-knowledge endeavours.
Society
- Someone will decide to leave agency to a set of AI or to agent-based frameworks, start an economic entity (trust? fund?) that does not necessarily do financial investment but starts delivering direct services to client and self-manages.
- Political attacks using deep-fakes-like technology become common.
- We will explore new ways of building trust and confidence between economic agents.
- Because production is massive, bottleneck will be human capacity to absorb. We start relying on agents to interface with the wider information world, and we put more weight on human-based relations where things are important.
- As a reaction, offline becomes attractive.
Infrastructure
- A critical failure of aging infrastructure will occur, demonstrating non-linearity of effects.
- NEOM competes for infrastructure professionals with Ukraine reconstruction
- AI impact on infrastructure - more datacenters, more energy, more chips required.
- More sustainability in infrastructure in general, less available skilled professionals.
- General effort to stabilize uncertainty (supply chain, climate, wages, ..)
Jobs
- Meeting ‘secretaries’ make meetings simpler: transcription allow for a full memory of exchanges. They can be queried in natural language.
- Automated translation, bidirectional, video-true, allow for better collaboration between areas. Barriers to competition appear to drop, only for cultures differences to become the next hurdle.
- First virtual ‘projects manager’ are mobilized on small/medium scale projects. They are equipped to err manage scope, budget and timelines, along with a keen understanding of contracts and smooth communication delivery. Humans pretty much required to check and approve content.
- ‘Buddies’ appear - when delivering a standard work, trainers equip buddy with an understanding of what your work means, how it’s delivered, and still manages to equip them with the knowledge of past reports.
- ‘TuringBots’ reshape the work of developers.
- We start to see ‘virtual colleagues’ - avatars based on one’s appearance, with access to past mails and chats, and who can take over when someone is on leave. Or organize meeting (timing, agenda, …). Or keeps a memory when someone leaves an organization or retires.
- As a reaction to creation of free content, “bullet point mails” to colleagues and possibly clients start to appear.
Future of work
- People get scared about what AI does. Society reacts by putting more controls - but individual entities keep on progressing due to lack of enforcement.
- Global anxiety raises in the knowledge sector.
A bit more on genAI
- After generalist genAI, dedicated models appear. Small ones are generated for specific purpose. Can be trained by content generated by larger ones.
- Smaller models rely heavily on good-enough, open-source models.
- As a consequence, agent-based frameworks grow (even more) on 2024.
- Dedicated hardware appears and allows running models on a phone.
- Supply chain security and auditability of one’s models appear as a reaction to models poisoning.
Security
- Cybersecurity threats still increases - starting with possibly seeing some proprietary models leaking. What happens when not a report leaks, but the tool capturing your business model?
- Information wars accelerate. The “free” production of content and models, in all media, saturates absorption. Trust is key, trusted curation occurs more.
- Privacy technology accelerates - (fully) homomorphic encryption starts to build.
- Trust, again, is key. A colleague calling you in full video for help might very be an external agent looking for information.
