2023-12-21 03:11:00
Unclear or no guidelines
Another result which unsurprisingly demonstrates this informal dimension: two thirds (66%) of Swiss employees are not trained for the appropriate use of generative AI, and just as many employees have not received training for secure use. . And if training is lacking, guidelines are also lacking: 77% of Swiss employees indicate that their company has not defined clear guidelines on the use of these tools on a daily basis. For 33% of them, these guidelines are completely lacking, and 41% indicate that their employer does not adopt a clear position on this subject.
Of course, the surge of ChaGPT was rapid. Enough to catch companies off guard, particularly those whose means are limited to respond to this upheaval.
But this data remains alarming when we know that generative AI is now part of the daily work of many employees: around two thirds (65%) of people questioned in Switzerland confirm that they are more productive thanks to AI, according to the study. But above all, no less than 61% of survey participants in Switzerland have already presented results generated by AI as their own. Within the Confederation, this figure is particularly high among respondents working in science and biotechnology (86%), as well as in the media and entertainment sector (81%).
An unmeasured risk of errors
This last point undoubtedly illustrates the heart of the problem. If many employees present results generated by AI as their own, it is undoubtedly also because the instructions are not clear or because the use of these tools is prohibited. The problem: these same employees are not trained in their use.
This reality legitimately raises fears of misuse of AI by certain employees – who might enter confidential data into unsuitable tools or who would not be able to identify certain limits or errors in content provided automatically. And colleagues and superiors are not aware that the results presented are those of AI, so many safeguards disappear.
Also read: Between directives and optimization: the ChatGPT puzzle in business
Ignoring the reality of increasing use of ChatGPT in business by employees is therefore not only counterproductive, it can also be dangerous. An extreme example, in Brazil, a judge used artificial intelligence to help him make his decision. Result: the sentence contained false references.
Even if the observation may seem self-serving, Salesforce does not say anything else in its press release: “To reap the benefits of generative AI, companies must establish clear guidelines for appropriate and safe use of the technology. In addition, employees must be trained accordingly and investment in reliable tools must be considered.”
Training courses, such as those from the ImpactIA foundation or independent experts, exist in Switzerland. Professionals can also help set guidelines. An investment in time and money today, but which might help avoid many disasters tomorrow.
Also read: Why and how businesses should train in ChatGPT
1703134022
#work #Companies #tune #employees