#GenetecConnectDX: Beware AI ‘False Promises,’ Warns Genetec President

Written by

Speaking in the keynote session of the Genetec Connect’DX digital conference Pierre Racz, president of Genetec, reflected on the false promises of Artificial Intelligence (AI), the state of the technology and its impact on our lives.

“Science and technology are morally neutral – how we use the fruits of science and technology is not,” he said. “Be skeptical of technology that works best when you need it the least.”

Racz explained that AI is effective if the job it’s required to do is very limited, claiming that AI is at it’s best when used to solve really narrow and well-defined problems.

However, Racz also highlighted the importance of understanding the limits of AI technology, adding that proper engineering of AI requires:

  • Knowing the limits
  • Avoiding wishful thinking
  • Being open and honest about the limits
  • Operating the technology within the limits
  • Making contingency plans for when technology fails
  • Taking into account the secondary and tertiary effects of the technology

Furthermore, Racz argued that real AI doesn’t actually exist: “Real stupidity exists; do not confuse the appearance of intelligence with actual intelligence.”

AI does not understand what we actually ask it to do, he said, for example: “When AI identifies a pedestrian in a picture it does so by recognizing lines and textures, that’s the only concept of a pedestrian it has. Nothing more.

“The big danger is not that AI will rebel against us and take over the world, it’s that it’ll do exactly what we ask it to do,” without the ability to add any additional logic.

AI is also susceptible to designed attacks, Racz pointed out. “It has short-term memory that can be exploited to produce unintended output,” AI algorithms are “generally bad at counting” and “just like humans, AI is susceptible to optical illusions.”

To conclude, Racz warned against misinterpreting “crafty [AI] guessing for intelligence or thinking” and warned against relying on false promises of what AI technology can achieve. Instead, he said, there is no substitute for human judgement and oversight. “Keep the human in the loop and the human can provide intuition and creativity while the machine does the heavy lifting.”

What’s hot on Infosecurity Magazine?