The limitations of artificial intelligence 2021
bigest limitations of artificial intelligence Ai
Limitations of AI
- Contant
- Ai can fall to
- Ai does not understand causal reasoning
- Lack of creativity
- High maintenance cost
More on Artificial Intelligence Though Al is seen by many as the next big thing, it still has severe limitations that may have unforeseen and potentially disastrous consequences if it is not implemented in the correct manner.
- AI can fail to
o Machines or software can fail anytime due to some bug or human error. For example, Microsoft was forced to disable its Tay chatbook, which was able to learn through its conversations with users, after Internet users managed to make it racist in less than one day. Similarly, Facebook was forced to admit when its bots reached a failure rate of 70% and started talking to each other in a strange language only they understood.
- AI does not understand causal reasoning
Al can only do what it is programmed to do. Granted, it normally does this extremely well. but unlike a real person it is unable to make split-second judgements, For example, any good marketer will change or cancel any scheduled messaging when a tragic event occurs. This is because we as humans are capable of showing empathy and compassion to the victims and their families. A machine, however, does not have the ability to show emotion, so could get a company in trouble in times of crisis if it's not carefully managed or controlled.
- Lack of Creativity
Creativity remains a vital component of a successful campaign. Machines simply lack the ability to be creative. Unlike machines, humans can think and feel, which often guides their decision making when it comes to being creative. When it comes to originality and creative thinking, a machine simply cannot compete with the human brain.
- High Maintenance cost Cost
is another key consideration with procuring Al technologies. Pains that lack i house skills or are unfamiliar with Af often have to outsource, which is where rule Cost and maintenance come in. Due to their complex nature, smart test expensive and you can incur further costs for repair and ongoing, malaterar UR. Tie computational cost for training data models etc. canalsobe an additiomalezpene
Comments