Three years ago, LeMonde.fr used, for the first time, a robot-editor to publish the election results. If it hadn’t been in the tech press, nobody would even have noticed. The quality was excellent, and the human eye could not perceive the difference between an automatic text or one written by a journalist.
For almost ten years now, part of the stock market news of some American newspapers and many websites have been written by robots.
This trend will intensify in the coming years.
Why write with a writing robot?
Content consumption has never been greater than it is today (2018). More and more people are connected to the Internet and spend even more time on it than ever before. Writing has opened to everyone and the areas of interest are expanding. There are thousands of subjects for which it is easier to read a text rather than a table of figures. Narrative Sciences specializes in writing board reports and converting Dashboard to text format. They confirmed their dominant position on the subject of NLG (Natural Language Generation) with a record $40 million fundraiser.
Finally, the advent of Vocal Search will require more structured content. Indeed, even if a vocal search does not necessarily require different content, the vocal answer, through the vocal assistants (Amazon Echo, or Google Home for example) will provide structured answers. A great majority will be completely written by a robot before being read by a system of text-to-speech.
According to the BBC, by 2020, 90% of fresh news content will be written by robots. The Associated Press in the United States already produces 10,000 economic and sports (baseball) reports each month.
Marketing content first and foremost.
We have seen an increase in marketing content for three years now. Storytelling and personalized marketing make content marketing a priority for decision-makers.
However, it is impossible for marketing teams to produce content for every user segment, every location, and every customer journey.
Automatic content generation is therefore a massive solution that complements journalistic writing. A robot is not, as it is, capable of storytelling but rather of writing structured and factual content.
So, what does the robot write after all?
Let’s look at a few examples from different industries to better understand the potential of a writer robot.
1. Construction sector
The construction sector is, by definition, a local market. It is therefore important to be able to write local content
| Entity name and contact details | Function | Entity country of establishment | Processing location(s) |
|---|---|---|---|
| OVH | Cache server | France | France |
| WALDC | Storage, databases | Belgium | Belgium |
| AWS | Storage, database | United States | Ireland |
2. Media and celebrities
In the media context, we find content that can be generated on the basis of an interpretation of the photo.
| Type of data | Processing | Storage time |
|---|---|---|
| IP addresses of end-users and of Verbolia users | Encrypted, anonymized then stored | Milliseconds before being encrypted, then 30 days |
| Email address of Verbolia users | Stored | As long as the user does not delete his/her account |
Another example in the sports world demonstrates the algorithm’s ability to link information together. In the case below, we find a description of two football players who share common elements of their biography.
| Service Level | Pro | Business | Enterprise |
|---|---|---|---|
| Service Availability | 99.5% | 99.5% | 99.9% |
3. Stock markets
As part of stock market, many U.S. sites write automatic reports based on market data and typical technical analysis indicators.
[table “5” not found /]
4. Politics and elections
[table “6” not found /]5. Real estate ads
Robots are very useful for writing content based on structured data. The world of real estate is a perfect example.
[table “7” not found /]
Is a writer robot better than the human?
Currently? It certainly is not. The generated content is based on a weak artificial intelligence. This means that robots are simply able to make a coherent, quality and complete text. However, they lack many of the qualities that journalists can have:
- Sense of humor
- Critical spirit
- Analytical and cross-checking spirit
The conclusion is that there are many cases where these qualities are not necessary, leaving room for human writers to produce quality content.
