This paper addresses the present state of play and future trends, uncertainties and possible disruptions of Artificial Intelligence (AI) and Big Data in the following areas:
Political: AI is biased, difficult to scrutinise and to estimate its power, and – more so when embodied in autonomous systems – potentially dangerous. Policy responses are accountability, transparency, safety and control, and public debate. These should be based on ethics. AI will lead to better governance, more debate, new policy actors and processes, a contest over centralisation, and the EU catching up. AI might progress in a revolutionary or evolutionary mode, lead to new political divisions, and change democracy. AI might be misused as a “superior orders” defence. What if data analysis changes or replaces democracy?
Socio-economic: Big Data is changing the role of data, is often dependent on sensitive information, is handicapped in the short term but better in the long term due to data protection, and its industry is in danger of monopolisation. AI lowers the cost of prediction, replaces human prediction and human labour and causes social problems, increased nudging and misuse of the term AI. AI will lead to more data, economic growth and more job market distortions. AI might lead to new industry giants, a request for more privacy, new state solutions, yet unknown jobs, AI taxes and increased state ownership. What if new economic ideologies emerge, singularity strikes or AI monopolies are broken up?
Geopolitical: AI is increasing the power competition between the US and China and gives both more power. Europe tries to create businesses and find its strengths. All are investing in military solutions and the west has a slight disadvantage here. AI will lead to a shakeup of the international system, hierarchies and networks becoming more powerful, and real-life deception being more difficult. AI might lead to China becoming the most powerful power overall and in AI. The future of AI R&D and the success of Europe’s broad approach is uncertain. What if there are two digital worlds, China becomes a data-privacy defender, and AI become targets?
Technological: Superintelligent AI is invested in and researched, challenged by philosophy, and possible this century. It might imitate the brain, be assembled together or be designed by other AIs. An intelligence explosion or a conscious AI could be possible, and might be the last invention of humanity. It would require long term funding, need to overcome many technical hurdles, be dangerous due to its intellect, possibly be contained with collective intelligence, and maybe have humans lose their jobs, safety or purpose.
Key questions for policy-makers: What makes European AI distinctive? What areas can and should we prioritise, if any? What should be regulated? How could and should the EU foster AI development, avoid monopolisation, provide data pools, use high data standards, link researchers and corporations, balance fundamental with applied AI research and private with state funding, boost applications, compensate for job loss, keep AIs away from dangerous actors, support EU foreign policy (neighbourhood, FPI, democracy and human rights, aid and development, economic freedom), improve our lives with AI, change the geopolitical AI race, deal with autonomous weapons and superintelligent AI and organise Foresight?