Listen to Einstein, he has ear of the boss
No one likes a snitch, especially one who sucks up to the boss. But executives who fear their colleagues will beat them up the greasy pole have a new rival – a robot who has the CEO’s ear.
Marc Benioff, founder of the American software company Salesforce, revealed that the company’s AI software, Einstein, has attended weekly senior management meetings for a year, criticising executives’ reports and ruffling feathers.
‘‘I have my top 30 or 40 executives around my table,’’ he said. ‘‘And we figure out how we are doing as we look at all of this analysis. But now I have a new person with me and it’s kind of an empty chair.’’ Benioff said that after all his human executives had spoken, he turned to Einstein for the final word.
Benioff said that, after one recent meeting, Einstein questioned the projections of a senior European employee. Einstein said: ‘‘Well, I don’t think this executive is going to make their number.’’ ‘‘I’m so sorry,’’ he added.
Benioff said that his human subordinate was visibly upset by the claim. Einstein, his apology notwithstanding, proceeded to break down his analysis in cold figures – much to the human’s embarrassment.
From a boss’s position, these brutal instant insights were hugely useful, Benioff claimed. ‘‘I could say, ’Well, you’ve got this problem, and this person hasn’t got that in place, and you didn’t do this over here.’ ‘‘
Salesforce wants to sell its Einstein AI to businesses for analytics and identifying sales leads. Analysts believe that robomanagers will be supervising millions of people to some degree within the next few years. Companies already use algorithms to evaluate aspects of staff performance and AI enthusiasts believe they will increasingly make management decisions.
For those fearful of roboschemers, there’s some temporary consolation. Experts stress that despite the ability of AI to solve specific problems and spot particular errors the technology is far from ‘‘thinking’’ in any meaningful sense.
Paul Armstrong, of Here/Forth, the technology consultant, said: ‘‘The near future may see robot ‘intelligence machines’ but managing goes way beyond simply making predictions and using an algorithm to come to a conclusion.
‘‘While the AI industry continues to leap forward, the days of true robot bosses are far in the distance. The technology and the interfaces simply aren’t there yet. The algorithms may get you the right answer but, for the main, they still need significant programming for high-level work. In the future, they won’t need this programming, which is when AI really gets interesting – and potentially dangerous.’’