Data protection and AI challenges
There are many challenges and uncertainties with the AI technology in relation to the data protection legislation.
Examples of challenges include:
The possibility of a so-called conversational leak of artificial intelligence
In the OpenAI User Terms and Conditions, individuals agree that the company can use the content that the person in question posts, to improve and develop the service. If that content contains personal data, then it can be assumed that the company can, and will use the content, for other purposes.
Creation of new personal data (information gap)
The artificial intelligence can create or exaggerate information, invent something, which doesn’t necessarily always turn out to be correct.
As an example, in California, a law professor was accused of sexual harassment, in reference to an article in The Washington Post. The article had been created by artificial intelligence - it did not exist in fact.
The right to be forgotten
Challenges are that according to the data protection legislation, individuals have the right to be forgotten.
However, as the examples have shown, grammar models can make mistakes. The law professor in the example above might wish to request that these incorrect information be deleted, but the fact is that it is not easy to delete or correct data in grammar models. Microsoft has described this as "as simple as taking one ingredient from a cake you have baked. That is not possible.
Principle of data protection law on transparency
Individuals have a right to information about the processing of their personal data.
However, it is clear that it is challenging to meet this principle when personal data are processed in conjunction with artificial intelligence programs. The technology is complex and it can be difficult to understand and explain why it "says" what it says. Even by those who have been involved in developing the program from the beginning.
The principle of proportionality of personal data
The function of the artificial intelligence is dependent on the massive amount of data
It has been stated that new large electricity plants need to be established for the sole purpose of powering electricity for so-called artificially intelligent data centers.
It is therefore clear that it is difficult to meet this principle to some extent with the use of artificial intelligence.
