Why AI Is Not Really Intelligent—And Why That Does Not Matter Either

Despite the great AI successes of recent years, we are still in an era of very

formal, machine AI. Figure shows that the underlying methods and

technologies have not fundamentally changed since the 1950s/1960s to

today. However, due to the increased amounts of data and computer capac-

ities, the methods could be applied more efficiently and successfully. The

so-called deep learning approaches brought about an immense leap in qual-

ity. These massive gradual improvements to “machine learning on drugs”

allow us to perceive a quasi-principle leap in AI that does not actually exist

in this way. The systems are still learning according to certain rules and set-

tings, patterns and distinctive features.

The next important step in the evolution of AI is the ability of the sys-

tems to learn autonomously and proactively to a wide extent. The first

promising learn-to-learn approaches were applied in the AlphaGo example

described. In addition, there are numerous promising research approaches

in this area that will lead to algorithms adapting themselves or that will also

develop new algorithms. This will, however, continue to happen in a rather

formal-mechanistic understanding. This has little to do with a human’s abil-

ity to learn. The next step of evolution, which then also contains human-like

abilities such as creativity, emotions and intuition, is a distant prospect and

eludes a reliable temporal prognosis.

From a business point of view, this discussion may appear to be academic

anyway. The decisive factor is the present-day perceived performance of the

AI systems. And even today, they outperform human performance in many

areas. Figure shows the development of AI performance in image recog-

nition. Even if the AI systems are still not perfect with their misclassifica-

tion of 3% today, they have been outperforming the classification skills of

humans since 2015. Thus, these systems can recognise the likes of reliable

cancer diagnoses, fraud detection or other relevant patterns. This also applies

to speech recognition.

Comments

Possible Limitations of AI-Based Bots

The examples above already show the present-day potential of AI-based bots. At present, these systems are still in an early stage and still have certain limitations and potentials for optimisation. Twitter Bot Tay by Microsoft Most bots at present are reactive service bots. Engagement bots that actively interact with the users as market and brand ambassadors go one step further. The most famous example here is the chatbot Tay by Microsoft. Microsoft removed Tay from the web apologetically within one day. The example shows that the uncontrolled training of bots by the community can lead to fatal consequences. AI systems still have to learn ethical standards. It thus becomes apparent that even bots require a kind of guideline. Like a journalist has to observe editorial guidelines, bots have to observe certain standards. The next generation of AI-based bots must control and create the possible room for communication. IBM Watson has been able to celebrate quite a few respectable resul...

Robot Journalism Is Becoming Creative

Algorithms are able to automatically search the Web for information, pool it and create a readable piece of writing. In addition, data-based reports in the area of sport, the weather or finances are already frequently created automat- ically today. Recently, for example, merely a few minutes after Apple had announced their latest quarterly figures, there was a report by the news agency Associated Press (AP): “Apple tops Street 1Q forecasts”. The financial report deals solely with the mere financial figures, without any human assistance whatsoever. Yet, AP was able to publish their report entirely via AI in line with the AP guidelines. For this purpose, AP launched their corresponding platform Wordsmith at the beginning of 2016, which automatically creates more than 3000 of such financial reports every quarter, and which are pub- lished fast and accurately. It is no longer that easy to distinguish between whether an algorithm or a human has written a text. Another exception of rece...

Why Use Machine Learning?

Consider how you would write a spam filter using traditional programming techniques (Figure 1-1): 1. First you would look at what spam typically looks like. You might notice that some words or phrases (such as “4U,” “credit card,” “free,” and “amazing”) tend to come up a lot in the subject. Perhaps you would also notice a few other patterns in the sender’s name, the email’s body, and so on. 2. You would write a detection algorithm for each of the patterns that you noticed, and your program would flag emails as spam if a number of these patterns are detected. 3. You would test your program, and repeat steps 1 and 2 until it is good enough. Since the problem is not trivial, your program will likely become a long list of com‐ plex rules—pretty hard to maintain. In contrast, a spam filter based on Machine Learning techniques automatically learns which words and phrases are good predictors of spam by detecting unusually fre‐ quent patterns of words in the spam examples compared to the ham e...

Tech & Fun

Search This Blog