A recent study suggests that up to 20 million manufacturing jobs could be replaced by robots around the world by 2030.  Predictions vary as to how quickly and how significantly AI and automation will impact the way we work.  However, there seems to be a general consensus that machines, particularly those driven by AI, will replace current jobs at an ever-increasing rate.  Gartner, for example, predict that by 2021, AI will replace 6.2 billion hours of worker productivity.

But it's not all doom and gloom.  Although worrying on the face of it, it presents an opportunity to retrain people and to employ human skill more effectively, where it is needed most.  The UK Government's Made Smarter Review, for example, predicts that up to 300,000 jobs will be created in the UK by 2025 by digitalisation.  The importance of initiatives like the new National Retraining Scheme is clear.

We're seeing a marked increase in technology and services solutions relying on AI or automation.  Often, we don't see the associated supply contracts developed accordingly.  Sometimes, that's fine.  However, where AI and machines replace human input in services arrangements, it's worth considering how this should be reflected in contractual provisions.

Some points to consider:

1.  Liability - Should contractual liability be approached in a new way?  With AI and automation, the risk of one-off errors is often low (as human error is minimised) but, if an error does arise, it is likely to be repeated and perpetuated.

2.  IP - Is the contract clear on who owns the IP in AI that has been developed based on your data?  This will be particularly important if the people that currently drive your services have been replaced by a machine.

3. Continuous Improvement - How precise should your provisions be and how should you deal with potentially drastic improvements and the impact of change?

4. Exit - As human input is replaced during the term of the contract, what support will be required on exit to handover to the next service provider?

5. TUPE - What impact will a steadily reducing level of human resource have on the relevance of TUPE provisions?

This is just a snapshot and there are, of course, other issues to consider (but a machine will probably soon be handling them too).